Comments on: SCD: Slowly Changing Dimensions or SchrÃ¶dinger’s Cat Dilemma?

By: Michael Amamoo

Michael Amamoo — Wed, 27 Jan 2010 06:10:16 +0000

Hi Oracle Alchemist, I like your comparision of quantum physics with slowly changing dimensions. Currently working on a project ” Data Warehouse Design and implementation to address the problem of slowly changing dimensions”. I am using SCD type 6 (1+2+3). Just to add that apart from external triggering events (e.g. change of address), data model changes can also trigger SCD.
Following a purist approach, one can (given storage space) capture all change states of an entity but to what end?
Thanks for pointing out that in the end, the most important change to capture is data that is of relevance from the business perspective.

Regards,

Michael Amamoo

By: Vijay

Vijay — Thu, 24 Jul 2008 10:02:50 +0000

Nice article. Clearly justifies the purpose of the article.
Moreover learnt something about quantum physics as well

By: TERRY FRASER

TERRY FRASER — Wed, 14 May 2008 15:36:50 +0000

Schrodingers cat scientificly speaking,cat tunneling with branes,

By: Pythian Group Blog » Log Buffer #40: a Carnival of the Vanities for DBAs

Pythian Group Blog » Log Buffer #40: a Carnival of the Vanities for DBAs — Fri, 13 Apr 2007 17:15:44 +0000

[...] Karam, the Oracle Alchemist, let the cat out of the bag (or out of the steel chamber perhaps) about indeterminacy in your Oracle databases. The crux of this excellent article is the question, “How can we record a fact when important [...]

By: The Oracle Alchemist

The Oracle Alchemist — Mon, 09 Apr 2007 14:11:32 +0000

Okay, granted. Perhaps I shouldn’t have used a numerical amount pertaining to the results as an example; however, the fact (no pun intended) still remains that slowly changing dimensions do exist, and require some resolution. Instead, I could have used the client’s last name or the location’s address as a changeable dimension value, as those could also cause problems with report output.

Had I filled in the model more (hey, I said it was incomplete) I might have had other examples that I could have used. I understand your point on the number of deductions, but SCD resolution is still a necessity for the reasons I mention in my article.

I actually had a nagging thought about the same thing you mention after looking at the post a couple times, but I wasn’t sure if I should change it once I had posted it. I appreciate you reading and commenting!

By: John Flack

John Flack — Mon, 09 Apr 2007 12:43:38 +0000

You’ve simply modeled this wrong. Number of dependents is not properly part of the Client dimension, it is a fact. As a fact, it has relationships with both the client dimension and the time dimension. Thus no paradox – client has 3 dependents associated with the 2006 tax year and 4 associated with the 2007 tax year.